Single site refractory effect

As shown with SSS in Figure 4, stimulation with two pulses on the same site at an IPI of 0.5 ms elicited an LFP response of similar size to that elicited by a single pulse.

The strong refractory effect of the first pulse on the activity to the second pulse can be better visualized by plotting the LFP area versus the different IPI values between 0.5 to 10 ms, during which only a single fused LFP exists. For IPI values greater than

58

about 10 ms, there are usually two separate LFP peaks. Figure 6 displays a typical SSS data set from one animal (for details on LFP area calculation, see Methods:

Data analysis). The LFP area for different IPI values between 0.5 to 10 ms (abscissa) and different levels (colored curves) are plotted. The ordinate corresponds to the normalized (Figure 6A) and normalized (Figure 6B) LFP area. The non-normalized curves show how the total LFP area increases with higher levels.

Generally, the area is low at an IPI of 0.5 ms and increases as the IPI increases up to about 2-3 ms. The IPI area then begins to decrease and reaches an approximate

plateau between about 5-10 ms.

Figure 4-6. Local field potential (LFP) areas for single-site stimulation (SSS)

A) LFP peak areas are plotted for different inter-pulse interval values across different stimulation levels in dB relative to 1 µA (for details on LFP area calculation, see Methods: Data analysis). These data correspond to SSS-2 in Table 1. B) Same curves as in A except they were normalized by dividing each value by two times the LFP area elicited by a single pulse. Therefore, a value of 0.5 indicates that the LFP area elicited by two pulses on the same site is equal to the LFP area elicited by a single pulse (i.e., full refractory for the second pulse). A value of 1 indicates that the LFP area to two pulses is equal to the linear sum of two individual areas to single pulses. The black dashed line corresponds to the averaged curve across all solid curves. The blue dashed line corresponds to a saturated, flat curve and is not included in the average (see Results: Single site refractory effect for further explanation for excluding curves for the averaged curve).

59

This general trend is observed across all levels, although the curves at very low or high levels appear flatter than those for the middle levels. At very low levels, the LFP activity is quite small relative to the background response fluctuations across IPI values, thus the trend is not clearly visible. At very high levels, we noticed that the LFP activity begins to reach saturation in which the second pulse does not contribute any additional activity to that of the first pulse even outside of the refractory period (>

2 ms). For the averaged curves presented in Figure 6B (black dashed line) and Figure 7, we only used the individual curves for levels that exhibited sufficient and non-saturating LFP activity. These levels are listed in Table 1 for SSS-1 to SSS-7.

Across levels and IPI values, there also appears to be some fluctuations in the LFP activity, likely related to anesthesia depth (i.e., A1 is sensitive to anesthesia level), stimulation effects over time (though stimuli were randomly presented across trials to minimize adaptation), and intrinsic neural oscillatory activity. These types of A1 fluctuations have also been observed for acoustic click stimulation with varying IPI values in cats (Eggermont and Smith, 1995) suggesting that it is not solely due to artificial electrical stimulation effects of the ICC. From our data, it is not possible to determine to what extent these fluctuations actually represent functional effects relating to specific IPI values. For the remainder of the paper, we will focus on the general shapes of these curves rather than the detailed fluctuation patterns across IPI values and levels.

To better visualize the extent of the single site refractory effect, we normalized the curves in Figure 6A as shown in Figure 6B. We divided the LFP area elicited by two pulses by two times the LFP area elicited by a single pulse. Thus a value of 0.5

60

indicates that the LFP area elicited by two pulses is equal to the LFP area elicited by a single pulse (i.e., full refractory for the second pulse). A value of 1 indicates that the LFP area to two pulses is equal to the linear sum of two individual areas to single pulses. It is clear that across all levels, the LFP area to two pulses with an IPI of 0.5 ms is approximately similar to the area elicited by a single pulse (normalized area equal to ~0.5) indicating full refractory when stimulated with the second pulse. As the IPI increases, all curves generally exhibit the same pattern in which the normalized area increases up to an IPI of about 2-3 ms and then decreases to an approximate plateau level by about 4-5 ms. The curve corresponding to 48 dB (dashed blue) is approximately flat due to saturation, thus was not included in the averaged curve (black dashed line). Based on these curves, there appears to exist some neural mechanism that enhances A1 LFP activity above the plateau level (about 0.73 for this example) for IPIs less than 4 ms but the LFP activity then begins to decrease at about 2 ms due to the refractory effect. If there was only a refractory effect without a neural enhancement mechanism, then we would expect that the normalized area would increase monotonically from 0.5 to the plateau level (without the hump at 2 ms) for increasing IPI values as the neurons recover from a refractory state.

Interestingly, the plateau level is not equal to 1, indicating that consecutive stimulation of a similar population of ICC neurons does not simply elicit a linear sum of LFP activity in A1. It is not clear whether this is due to a subset of neurons that fully remain in a refractory or suppressed state for a period longer than 10 ms or related to some nonlinear way in how charge accumulates around the recording site.

Nevertheless, Figure 6 clearly reveals an enhanced activation region for IPI values between about 2-4 ms that is reduced by some refractory effect at shorter IPI values.

61

As will be shown in the next section, MSS overcomes this refractory effect by activating different neural populations in which the normalized area continues to increase with ms), plateau level (between about 0.58 and 0.74), extent of enhancement above the plateau level (between 4 and 32%), and starting point of the plateau portion (between 4 and 6 ms). In fact, the two bottom curves (SSS-1 and SSS-4) did not exhibit any noticeable enhancement effect but gradually recovered from the refractory state to the plateau level by about 4 ms. Nevertheless, they all exhibit some general trends in which a strong refractory effect to the first pulse on the second pulse response Figure 6B (dashed black line) for all SSS data sets listed in Table 1.

62

that we observed consistent trends across all our different ICC locations and levels indicating that the trends observed in Figure 7 appear to reflect general properties of ICC neurons. In other words, stimulation with a low level will activate a small volume of ICC neurons whereas a higher level will activate a larger volume of neurons. Yet across levels we generally observed similar trends and this was also true for different ICC locations across animals. In a future study, we will systematically stimulate different locations across a given ICC lamina to assess if the differences observed across curves in Figure 7 are associated with different subregions throughout the ICC and specific coding properties. It is also possible that those differences correspond to different recording locations within an isofrequency band of A1.